Collision Callback Information

[smashthewindow]

void _stdcall EntityCollisionCallback( TEntity entity0, TEntity entity1, byte* position, byte* normal, byte* force, flt speed )

Is the position, normal, force parameters relative to entity0 or entity1?

[Canardia]

I think it's neither. It's the global position of the collision itself. For example if two spheres collide, it would be a point on both sphere's hull.

[smashthewindow]

I think it's neither. It's the global position of the collision itself. For example if two spheres collide, it would be a point on both sphere's hull.

 

But then how would the normal or force vector be expressed in global coordinate?

Also is there a LE function to change position relative to entity0?

[Canardia]

There's no need for that, because when the collision occurs, you can always check the position of entity0 or entity1.

[smashthewindow]

I'm thinking of a teleport function which moves an entity with exactly the same movement info and relative position to the entrance as it came in, which is why I need the normals and the force.

Does the movement info ( velocity, force) get preserved after PositionEntity()?

I don't have access to a Windows platform till next weekend so unfortunately I cannot test this...

[Josh]

The position and normal are in global space.

 

There is no way to reposition a physics body without losing one frame of physics interactions, but you are unlikely to run into trouble. I would just get the velocity and angular velocity, then reposition the body, then set the velocities to what they just were. I would probably not do this in the collision callback, just set a value and then reposition the body after UpdatePhysics is finished, back in the main program body.

[Canardia]

But yeah, physics preserving teleports are possible and easy to do. I already did them in the LCP1 game years ago.